6. Results and discussion
6.1. Potentiometric immunoassays
6.1.1. Potentiometric enzyme immunoassay
55
56
means having a p𝐾𝑎smaller by at least 2 logarithmic units than the working pH.
Another alternative to keep the product in anionic form, is to stop the enzymatic reaction by concentrated NaOH, however this works only for endpoint-type detection.
(3) The commercial availability, (4) low toxicity, and
(5) chemical stability of both the enzyme substrate and the product anion is essential as well.
(6) Preferably the organic anion should form with high turnover rate (7) in a one-step enzymatic reaction.
While HRP provides limited flexibility in terms of anion-generating substrates, there is an excess of commercially available substrates for ALP and GAL enzymes. Remarkably, the anionic products generated by these two hydrolysing enzymes are practically the same since the artificial substrates differ only in having either an inorganic phosphate group or a β-D-galactopyranoside to be cleaved. The use of ALP, however, was immediately ruled out since the substrate also is an anion and there is little difference in the lipophilicities of the substrate and the product.
Figure 20 Scheme of the galactosidase catalysed reaction of methylumbelliferyl-β-D-galactopiranoside (DIFMUG) generating the ionisable product 6,8-difluoro-4-methylumbelliferon (DIFMU)
One substrate of galactosidase, i.e. 6,8-difluoro-4-methylumbelliferyl-β-D-galactopiranoside (DIFMUG) was found to fulfil all the above mentioned criteria. Although GAL is the least popular of the three most common enzyme labels in ELISA owing to its higher molecular weight (464 kDa), it is widely utilized, has a high turnover number, good compatibility with an extremely large range of buffers, and unlike HRP and ALP it is not present in the mammalian tissue, which helps to lower the chance of non-specific signals.
57 GAL has its pH optimum between 7 and 8, needs 1 mM Mg2+ for its activity as a cofactor, and can be inhibited by several metal ions, e.g. by Cu2+. The enzymatic reaction of DIFMUG, as shown in Figure 20, yields in the ionisable hydrolysis product 6,8-difluoro-4-methylumbelliferone (DIFMU), which has a p𝐾𝑎 of 4.9 and a calculated log 𝑃 of 1.83.
Table 2 Potentiometric selectivity coefficients (𝑙𝑜𝑔 𝐾𝐷𝐼𝐹𝑀𝑈−,𝑗
𝑝𝑜𝑡 ) of various membranes, measured by the separate solution method at 1 mM level, for detecting DIFMU. In the case of 4-methylumbelliferone anion (MU-) and p-nitrophenolate (PNP-) the pH was adjusted with concentrated NaOH, to *pH 9.0 and **pH 10.0, respectively, to ensure the dominance of the anionic form.
Plasticizer DOS o-NPOE o-NPOE
TDMA-NO3, mmol/kg 20 20 10
SO42- -4.9 -5.5 -5.9
OH- -4.0 -5.0 -4.5
Cl- -3.6 -4.0 -3.9
NO3- -1.9 -1.8 -1.8
MU-* -0.6 -0.6 -0.5
PNP-** -0.5 -0.5 -0.5
When using an anion-exchanger-based membrane for potentiometric detection in a heterogeneous ELISA immunoassay, the main interference to be expected arise from the inorganic anions of the working buffer solution. This buffer should ensure the optimal pH and ionic conditions for the galactosidase enzyme reaction. Potentiometric selectivity coefficients for the commonly encountered inorganic anions were measured by the separate solution method at 1 mM level, 𝑙𝑜𝑔 𝐾𝑖,𝑗𝑝𝑜𝑡 (𝑖 = DIFMU-) values for the different membrane constitutions tested are summarized in Table 2.
A compromise between the selectivity and the suitability of the buffer solution in terms of optimal enzyme function resulted in the use of 1 mM phosphate buffer containing 1 mM MgSO4 at pH 7.7 as working buffer, and the use of an ISE membrane composition of 40 wt% PVC, 60 wt% o-NPOE as plasticizer, and 10 mmol/kg TDMA-NO3 as anion-exchanger.
To obtain the potentiometric detection of PSA immunoassay in microtiter plates miniaturized liquid junction electrodes were prepared. The conventional, but miniaturized reference electrode was surrounded by 5 anion-exchanger-based working electrodes, as shown in Figure 12 (page 41), allowing the performance comparison of 5 electrodes simultaneously. The electrodes were first tested in 5 ml working buffer, than the DIFMU- calibration curves were
58
recorded in 150 µl volume in the microtiter plate wells. The electrodes gave Nernstian response with a detection limit of 8 × 10-5 M. This LDL was in good accordance with the theoretical detection limit determined by the selectivity coefficients (Figure 21 black). This indicates that no contamination or adsorption occurs during the measurement. Since a classic sandwich ELISA is an endpoint-type assay, i.e. the DIFMU- detection is performed after the enzymatic reaction is terminated with a stop reagent, calibrations were done in the presence of both 1 mM NaOH (Figure 21 blue) and 2 mM CuSO4 (Figure 21 red). (According to our preliminary measurement, when incubating GAL with 2 mM CuSO4 the enzyme activity drops to only 0.9 % of the initial value.) Although the use of NaOH seems to be advantageous in terms of keeping the product in anionic form via the resulting pH, OH- is less discriminated than SO42-, and it led to sub-Nernstian response slopes (possibly due to anion interference) and much higher LDL. In contrast, the addition of CuSO4 resulted in a linear super-Nernstian response of DIFMU- (~105 mV/decade), which has proved to be reproducible, and which considerably increased the sensitivity of the determination.
-7 -6 -5 -4 -3
-250 -200 -150 -100 -50 0 50 100
EMF / mV
log [DIFMU-]
59 mV
Figure 21 Calibration curves for DIFMU- in microtiter plates, i.e. in 150 µl sample volume:
(black) in the working buffer (1 mM phosphate buffer with 1 mM MgSO4 at pH7.7), (blue) in the same working buffer but with 1 mM NaOH and (red) in the same working buffer but with 2 mM CuSO4.
The determination of human PSA was performed according to the scheme shown in Figure 18 (page 49). The microtiter plates were modified with capture antibody having an affinity constant of 1.0 × 1010 M-1 [243], to which the PSA bound from human serum samples. After successive incubation in biotinylated tracer antibody, streptavidin-GAL conjugate and 0.5 mM DIFMUG, the
59 enzymatic reaction was stopped by 2 mM CuSO4, and the captured PSA was detected potentiometrically by measuring the generated DIFMU- anion in the solution. Each concentration was measured in 3 replicate wells, with 3-5 working electrodes.
All the assay steps were optimized in order to use the lowest concentration of the bioreagents providing maximal or close to maximal response, and thereby to lower the non-specific adsorption. The lowest concentration needed of the capture antibody, the tracer antibody and the streptavidin-GAL conjugate, was determined one-by-one by potentiometric DIFMU- detection, using 6-7 different concentration of the respective reagent, three parallel from each, and keeping the rest at the maximum (e.g. 50 ng/ml PSA, 20 µg/ml enzyme conjugate) or at the already optimized level. The ideal concentration was found to be 5 µg/ml for all the three components (for diagrams see Paper I).
BSA PF SB 1h SB 10min
0 5 10 15 20
blank response / mV
WB H2O
A
BSA PF SB 1h SB 10min
0 10 20 30 40 50 60
70
B
assay range / mV
WB H2O
Figure 22 Reduction of non-specific adsorption for PSA detection in potentiometric ELISA, in terms of (A) minimizing the potential response for the blank sample and (B) maximizing the potential range of the assay. Different blocking methods: 20 µg/ml BSA in the working buffer, Protein-Free blocking buffer (PF) and SuperBlock blocking buffer (SB) for 1 h and for 10 min were tested; and distilled water (H2O) and washing buffer (WB) was used in the washing steps.
The assay was also optimized in terms of reducing non-specific adsorption with a number of blocking agents (Protein-Free and SuperBlock blocking buffers, and 20 µg/ml BSA in the working buffer) and two different washing methods (distilled water and washing buffer). The goal was to minimize both the potential response of the blank sample, i.e. the potential difference between the 0.5 mM DIFMUG solution without enzymatic treatment and the 0 ng/ml PSA well (Figure 22A), and to maximize the potential range of the assay, i.e. the potential difference between the 0 ng/ml and 50 ng/ml PSA samples (Figure 22B). While the use of washing buffer minimalizes the extent of non-specific
60
adsorption, the potential range increases spectacularly when blocking with 20 µg/ml BSA solution.
0 10 20 30 40 50
0.0 0.2 0.4 0.6 0.8 1.0
A
[hPSA] / ng ml-1
A
0.1 1 10 100
-110 -100 -90 -80 -70 -60 -50 -40 -30 -20
EMF / mV
[h PSA] / ng ml-1
B
Figure 23 Calibration curves for PSA in human serum background using ELISA assays with (A) optical absorbance and (B) potentiometric detection.
The potentiometric ELISA detection of PSA from human serum gave linear semilog calibration in the range of 0.1‒50 ng/ml, as shown in Figure 23B. It provides a sufficiently low detection limit (≤0.1 ng/ml), which complies with the requirements of in vitro diagnostic PSA assays. The detection limit of the assay is determined by the selectivity of the electrodes, so further enhancements in this matter could improve the performance of the anion-exchanger based potentiometric immunoassays. When comparing these results with the conventional optical detection of the same assay using HRP labelling (Figure 23A), there was no significant difference at the 95 % confidence level between the two methods. Error propagation calculations, after fitting the experimental data with dose-response function, show that the uncertainty of the optical measurement in terms of concentration in the middle of the measuring range is 28.1 %, while the potentiometric assay has a relative error of 8.9 %. Although the procedure cannot compete in terms of analysis time with the highly parallel microtiter plate readers, it is by far superior to the previously published nanoparticle-based potentiometric immunoassays[203] [237]. The results confirm the applicability of anion-exchanger based potentiometric detection in diagnostic PSA assays.
61 Outlook
Following this work, recently, several publications by the group of Wei Qin utilized the same concept, the potentiometric detection of enzymatically generated lipophilic molecules with a simple ion-exchanger-based polymeric membrane. First it was shown[244], that reactive cationic intermediates rather than stable reactants or products can also induce large potential responses on appropriately formulated membranes with a cation exchanger. In many cases both substrates and products of enzymatic reactions are non-ionic, so these reactions are not considered in designing potentiometric biosensing schemes.
However, reaction intermediates having different p𝐾𝑎 values from substrates or products can be transferred into the membrane from the aqueous phase with the right ion-exchanger. This method was used for monitoring enzymatic reactions of HRP, and the peroxidase mimetic G-quadruplex DNAzymes[245]
[246] [247]. This way horseradish peroxidase has been detected with a detection limit at least two orders of magnitude lower than those obtained by spectrophotometric techniques. Different HRP substrates were utilized for the detection, such as phenol[245], o-phenylenediamine[246] and even the widely used TMB (3,3’,5,5’-tetramethylbenzidine)[247].
62